Redditor fine-tunes SDXL on 60 childhood photos to simulate memory recall
uisato · reddit · 2026-09-04
Reddit user uisato fine-tuned SDXL on 60 childhood family photos as a speculative study: instead of faithfully reconstructing images, the model produces unstable variations of spaces, faces and fragments that feel familiar but never existed — treating generative hallucination as an analogue for episodic memory as a reconstructive process.
The author stresses it's not "just a prompt": the pipeline combines Kohya fine-tuning, an audio-reactive geometry system built in TouchDesigner, and a custom rebuild of WarpFusion to interweave the geometries with the fine-tuned model, finished in Premiere, After Effects and Ableton Live.
Project files and tutorials are promised via the author's YouTube, Instagram and Patreon.
More from Multimodal
- Voxtral 3 ships a LiveKit-based voice agent SDK streaming listen-think-speak — MistralAI · 2026-10-21
- Voxtral Mini TTS 3 hits SOTA expressiveness in 35+ languages with tag-based control — MistralAI · 2026-10-21
- Mistral launches Voxtral 3 speech models and new Voice Agents API — MistralAI · 2026-10-21
- Open-sourced pixel art normals generator ships as Rust CLI and wasm library — HugoDuprez · 2026-10-20
- Fibo 2.0 is strongest for ecommerce and creator content, weakest at UIUX design — ArtificialAnlys · 2026-10-20
- Fibo 2.0 leads old Fibo on all 9 capabilities, best at reasoning and anatomy — ArtificialAnlys · 2026-10-20